The Use of (Q)SAR Methods in the Context of REACH

DeepAmes: A deep learning-powered Ames test predictive model with potential for regulatory application

The Ames assay is required by the regulatory agencies worldwide to assess the mutagenic potential risk of consumer products. As well as this in vitro assay, in silico approaches have been widely used to predict Ames test results as outlined in the International Council for Harmonization (ICH) guidelines. Building on this in silico approach, here we describe DeepAmes, a high performance and robust model developed with a novel deep learning (DL) approach for potential utility in regulatory science. DeepAmes was developed with a large and consistent Ames dataset (>10,000 compounds) and was compared with other five standard Machine Learning (ML) methods. Using a test set of 1,543 compounds, DeepAmes was the best performer in predicting the outcome of Ames assay. In addition, DeepAmes yielded the best and most stable performance up to when compounds were >30% outside of the applicability domain (AD). Regarding the potential for regulatory application, a revised version of DeepAmes with a much-improved sensitivity of 0.87 from 0.47. In conclusion, DeepAmes provides a DL-powered Ames test predictive model for predicting the results of Ames tests; with its defined AD and clear context of use, DeepAmes has potential for utility in regulatory application.

Multi-Strategy Assessment of Different Uses of QSAR under REACH Analysis of Alternatives to Advance Information Transparency

Under the Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) analysis of alternatives (AoA) process, quantitative structure–activity relationship (QSAR) models play an important role in expanding information gathering and organizing frameworks. Increasingly recognized as an alternative to testing under registration. QSARs have become a relevant tool in bridging data gaps and supporting weight of evidence (WoE) when assessing alternative substances. Additionally, QSARs are growing in importance in integrated testing strategies (ITS). For example, the REACH ITS framework for specific endpoints directs registrants to consider non-testing results, including QSAR predictions, when deciding if further animal testing is needed. Despite the raised profile of QSARs in these frameworks, a gap exists in the evaluation of QSAR use and QSAR documentation under authorization. An assessment of the different uses (e.g., WoE and ITS) in which QSAR predictions play a role in evidence gathering and organizing remains unaddressed for AoA. This study approached the disparity in information for QSAR predictions by conducting a substantive review of 24 AoA through May 2017, which contained higher-tier endpoints under REACH. Understanding the manner in which applicants manage QSAR prediction information in AoA and assessing their potential within ITS will be valuable in promoting regulatory use of QSARs and building out future platforms in the face of rapidly evolving technology while advancing information transparency.

QSAR Use in REACH Analyses of Alternatives to Predict Human Health and Environmental Toxicity of Alternative Chemical Substances

In 2006, the European Union (EU) enacted the Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) to address growing concerns of hazardous chemicals in the EU market. Under REACH, companies seeking authorization to use priority substances identified as substances of very high concern (SVHCs) and included in the authorization list must apply and submit health and environmental effects data in analyses of alternatives (AoAs) to the European Chemicals Agency (ECHA). To assess safer alternatives, especially in AoA hazard assessment cases where vital information could be missing or insufficient, quantitative structure-activity relationship (QSAR) nontesting methods have gained increasing acceptance and importance. This article assesses AoA applicants' use of QSAR sources and documentation while looking for meaningful trends. In this assessment, usage and frequency of QSAR sources were evaluated in 189 analyses of alternatives for 15 physicochemical properties and 19 human health and environmental endpoints to determine the scope of purpose of QSAR use in AoAs. We found that only 24 out of 189 applications cited QSAR sources to rank or evaluate the safety of their alternative substances relative to the REACH Annex XIV chemical. For human health and environmental hazard endpoints, applicants cited the Danish (Q)SAR Database (n = 63) and unidentified QSARs (n = 36) most frequently. While QSARs were not used to eliminate an alternative, 7.9% and 1.4% per maximum opportunity (MOP) of hazard endpoint and physicochemical QSAR predictions reported background information on alternatives using weight of evidence (WoE). In addition, 3.0% per MOP of hazard endpoint QSAR predictions supported the safety of the alternative while 0.7% per MOP of physicochemical QSAR predictions gave mixed support for their alternative's safety. Documentation regarding QSARs was absent in all 24 AoAs that used QSARs. Limited QSAR use and missing documentation may be the result of several factors, including inconsistent regulatory guidance. Integr Environ Assess Manag 2020;16:745-760. © 2020 SETAC.

Chemoinformatics: Achievements and Challenges, a Personal View

Chemoinformatics provides computer methods for learning from chemical data and for modeling tasks a chemist is facing. The field has evolved in the past 50 years and has substantially shaped how chemical research is performed by providing access to chemical information on a scale unattainable by traditional methods. Many physical, chemical and biological data have been predicted from structural data. For the early phases of drug design, methods have been developed that are used in all major pharmaceutical companies. However, all domains of chemistry can benefit from chemoinformatics methods; many areas that are not yet well developed, but could substantially gain from the use of chemoinformatics methods. The quality of data is of crucial importance for successful results. Computer-assisted structure elucidation and computer-assisted synthesis design have been attempted in the early years of chemoinformatics. Because of the importance of these fields to the chemist, new approaches should be made with better hardware and software techniques. Society's concern about the impact of chemicals on human health and the environment could be met by the development of methods for toxicity prediction and risk assessment. In conjunction with bioinformatics, our understanding of the events in living organisms could be deepened and, thus, novel strategies for curing diseases developed. With so many challenging tasks awaiting solutions, the future is bright for chemoinformatics.

RRegrs: an R package for computer-aided model selection with multiple regression models

Background

Predictive regression models can be created with many different modelling approaches. Choices need to be made for data set splitting, cross-validation methods, specific regression parameters and best model criteria, as they all affect the accuracy and efficiency of the produced predictive models, and therefore, raising model reproducibility and comparison issues. Cheminformatics and bioinformatics are extensively using predictive modelling and exhibit a need for standardization of these methodologies in order to assist model selection and speed up the process of predictive model development. A tool accessible to all users, irrespectively of their statistical knowledge, would be valuable if it tests several simple and complex regression models and validation schemes, produce unified reports, and offer the option to be integrated into more extensive studies. Additionally, such methodology should be implemented as a free programming package, in order to be continuously adapted and redistributed by others.

Results

We propose an integrated framework for creating multiple regression models, called RRegrs. The tool offers the option of ten simple and complex regression methods combined with repeated 10-fold and leave-one-out cross-validation. Methods include Multiple Linear regression, Generalized Linear Model with Stepwise Feature Selection, Partial Least Squares regression, Lasso regression, and Support Vector Machines Recursive Feature Elimination. The new framework is an automated fully validated procedure which produces standardized reports to quickly oversee the impact of choices in modelling algorithms and assess the model and cross-validation results. The methodology was implemented as an open source R package, available at https://www.github.com/enanomapper/RRegrs, by reusing and extending on the caret package.

Conclusion

The universality of the new methodology is demonstrated using five standard data sets from different scientific fields. Its efficiency in cheminformatics and QSAR modelling is shown with three use cases: proteomics data for surface-modified gold nanoparticles, nano-metal oxides descriptor data, and molecular descriptors for acute aquatic toxicity data. The results show that for all data sets RRegrs reports models with equal or better performance for both training and test sets than those reported in the original publications. Its good performance as well as its adaptability in terms of parameter optimization could make RRegrs a popular framework to assist the initial exploration of predictive models, and with that, the design of more comprehensive in silico screening applications.Graphical abstract

Robustness in experimental design: A study on the reliability of selection approaches

The quality criteria for experimental design approaches in chemoinformatics are numerous. Not only the error performance of a model resulting from the selected compounds is of importance, but also reliability, consistency, stability and robustness against small variations in the dataset or structurally diverse compounds. We developed a new stepwise, adaptive approach, DescRep, combining an iteratively refined descriptor selection with a sampling based on the putatively most representative compounds. A comparison of the proposed strategy was based on statistical performance of models derived from such a selection to those derived by other popular and frequently used approaches, such as the Kennard-Stone algorithm or the most descriptive compound selection. We used three datasets to carry out a statistical evaluation of the performance, reliability and robustness of the resulting models. Our results indicate that stepwise and adaptive approaches have a better adaptability to changes within a dataset and that this adaptability results in a better error performance and stability of the resulting models.

From descriptors to predicted properties: experimental design by using applicability domain estimation

The importance of reliable methods for representative sub-sampling in terms of experimental design and risk assessment within the European Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) system is crucial. We developed experimental design approaches, by utilising predicted properties and the 'distance to model' parameter, to estimate the benefits of certain compounds to the quality of a resulting model. A statistical evaluation of four regression data sets and one classification data set showed that the adaptive concept of iteratively refining the representation of the chemical space contributes to a more efficient and more reliable selection in comparison to traditional approaches. The evaluation of compounds with regard to the uncertainty and the correlation of prediction is beneficial, and in particular, for regression data sets of sufficient size, whereas the use of predicted properties to define the chemical space is beneficial for classification models.

Computational science in drug metabolism and toxicology

Computational scientific tools involving construction and testing of models, screening and data mining for drug and chemical induced toxicities and metabolism have significantly grown in experimental use to help guide product development and assist by enhancing certain areas of regulatory decision making. This themed issue of the journal entitled Computational Science in Drug Metabolism & Toxicology contains state-of-the-art review articles and perspectives covering a diversity of in silico approaches. Computational science tools have a strong potential for expediting our further understanding of drug metabolism and toxicity and are continually being developed and validated. The reader will gain an understanding of the current state of in silico tools and modeling approaches aimed at reducing these liabilities. In addition, how these tools are tested and developed for use in drug safety to support drug development efforts and a review of how they are used to predict genotoxic liabilities are covered in this issue. Computational science tools when properly validated and used judiciously can lend themselves as enablers to support drug safety assessment in investigative and applied settings.

Comparative PBT screening using (Q)SAR tools within REACH legislation

Small to medium sized enterprises (SMEs) in the EU are facing challenges due to the introduction of new legislation designed to protect consumers and the environment, REACH (Registration, Evaluation, Authorisation and Restriction of CHemicals). There can be high costs associated with implementing REACH because data on mammalian toxicity, environmental toxicity and environmental fate properties is required and if this data is obtained experimentally the cost is significant. These costs can be reduced if reliable quantitative structure-activity relationships ((Q)SAR) models are instead used to obtain the required information. In this paper we investigate how easily freely available (Q)SAR models can be applied for persistent, bioaccumulative and toxic (PBT) screening of 17 chemicals of interest to SMEs. In this study the PBT predictions obtained from the more user-friendly PBT Profiler and the Danish(Q)SAR database for the chemicals were compared with the results taken directly from the EPI Suite software. It was found that these widely used (Q)SAR databases might have some errors and examples are provided. It was concluded that extra care must be taken when considering the use of these databases for PBT screening. In addition, to increase the likelihood of a correct prediction, data estimates from various (Q)SAR models relevant to the PBT endpoints must be compared.